task Plex Media Serv:25382 blocked for more than 120 seconds?

server-linux

#1

Just moved my instance over to ubuntu 16.04 server. About every few hours, the server hangs with:
task Plex Media Serv:25382 blocked for more than 120 seconds
Tainted: G OE 4.4.0-62-generic #83-Ubuntu

PlexPy reports server down, and stopping the service takes 3+ mins. Starting it will only solve the issue for 30+ minutes. A full reboot is needed.

I have CIFS shares mounted on a nearby machine. It is all connected via LAN.
The Linux box is a VM, using 4 cores, x64 12G ram.

Logs in link, thanks in advance for any help.


#2


Here is a screen cap showing the errors, on the console.


#3

There's something very odd about your installation somewhere. Time to start digging. PMS behaves very well on Ubuntu in VirtualBox .

Setting up Ubuntu on VBox is straight forward. I typically grant it 128 MB of video ram for the desktop, 4GB of main memory, and a 32 GB virtual HD. Most importantly, I configure networking as Bridged. It stands up on the LAN as another host. Putting it in NAT adds another layer of translation and will always cause problems when PMS -> NAT -> your modem/router -> NAT -> Internet.

If you can get into the box otherwise with PMS stopped, Create a ZIP file of the Logs directory found under /var/lib/plexmediaserver/Library/Application Support/Plex Media Server.

Attach that zip file here.

First step is seeing if PMS is complaining at anything.


#4

Logs enclosed.


#5


Here is the setup of the VM.


#6

It's hanging for a bunch of reasons:

  1. Turn off IPv6 inside PMS and within the VM. Nothing at Plex.tv uses it. It is provided only to support local clients which are only IPv6 devices. Everything external to you is IPv4 because the internet is IPv4-only.
  2. It appears your machine is hanging up because of this.. Curl error 52 and 56. 56 is a failed SSL connection. 52 is the secondary empty reply.

Please do this, restart the VM as well and see what the next round brings to the surface.


#7

Disabled ipv6 and rebooted the VM. Fingers crossed.

Thanks for your help.


#8

Crashed again, Updated logs enclosed.


#9

Same issue came back again. :(


#10

@Stramondroid@gmail.com said:
Crashed again, Updated logs enclosed.

I have been looking at the last log which covered 2nd June from 14:14 to 16:23 - I need to know if these were the hangs that you noticed

I can see a few requests to the server not getting completed.
There was for xml for a photo at request at 15:10:34
and 9 requests from an android device (cellular) for user BmxWylie between 15:32 and 15:34 and also at through remote Plex Web at 15:57

Are these the hangs you are referring to?

On next occurrence would like you to get the output from this url on the server

http://127.0.0.1:32400/connections?X-Plex-Token=xxxxxxxxxxxxxxxxxxx
You will need to find the server token (should be in Preferences.xml in PlexOnlineToken or using plex web and look at the media info for a library item and it will be in the url

So start with a reboot so that the log covers the whole period and when you get the problem, obtain the connections response and paste here together with fresh logs

There is some deadlock - but normally deadlocks do not resolve themselves after a couple of minutes as you indicated.


#11

@"Stramondroid@gmail.com" to add to my earlier feedback, with regards to the 120 second hangs, the last log which covered 14:14 to 16:23 on 2nd June, there were no gaps in the timestamps in the logged messages - so not sure if there are two separate issues here. I could only see the requests that got stuck at the times i mentioned. Also there was continuous 1 minute remote access connectivity check being made (presumably by PlexPy) - it was every minute at xx:29 seconds on the dot without fail throughout the period.

And now looking at the earlier logs,
Server was launched at Jun 02, 2017 13:40 and shutdown at 14:12
and the logs before that
Server was launched June 02, 2017 13:35 and a crash may be at 13:37
and the logs before that
Server was launched at June 02, 2017 12:59 and shutdown at 13:27

What we had at 13:37 is loads of NULL characters written to the end of the log file
This has been seen to arise if the Plex area is mounted via NFS where there are issues with file-locking


#12

I will get this information over to you as soon as I am able. Thank you for taking a look at this.


#13

Release 1.8.0 of Plex Media Server has just been released for Plex Pass. It has the DASH transcoder deadlock fix

If you are a Plex Pass user, please upgrade and report back if this fixed your deadlocks and error 503. If not then we will need a new investigation with new set of diagnostics to investigate

See Release Notice http://forums.plex.tv/discussion/comment/1488006#Comment_1488006


#14

@sa2000 1.8.0 helped a bit, but I’m still seeing deadlocks on Debian occasionally on playback start especially from Apple TV and iOS devices.